I hit a bit of a thought roadblock with fanconcert so I decided to pull the editing parts out of it and make a generalized version, which I'd like to release soon so people can bash on it.

In a way, it's a bit of an idealized system as well. fanconcert has been refactored so many times and has enough data already that it's losing its agility. It's interesting to get to that point and examine why and use those lessons to not get there the next time, but I digress.

Problems with fanconcert

Problem: After almost a year of deployed use, fanconcert's limitations are showing. In some cases these limitations are intentional simplicities, in other cases they are just inimplemented features.

A glaring example is how labour-intensive inputting data is. People just don't like to do this. After posting over 4000 concerts myself I don't have much of a problem seeing why. It's just boring -- it's a lot like work!

Solution? Let computers submit information. How? Let my users create bots that crawl the Internet, parse the pages to extract the data and submit the data to fanconcert. This strategy has a few benefits:

1. People love hacking up tools, if you open up your software to let them do it.
2. The data entered by bots can still be corrected and added to by humans but a majority of the grunt work will be done.
3. If bots are refreshing regularly from sources, data will be updated faster than if done by a human, in some cases.

Problem: fanconcert is not simple enough!

Solution? Users fall into two major camps: people that just want to be updated with information and those that are also willing to submit and correct information. The great websites out there do a good job of separating viewing from submitting in order to make plain viewing easier. I really like this strategy and I'm going to try harder to do this with fanconcert.

Problem: fanconcert doesn't have enough users.

Solution? This is a chicken and egg kind of problem: more users will come when there's more information to read, there will be more information to read when there are more users. Bots could help this situation by increasing the amount of information without needing more users to enter the information. That's why I'm heading in that direction...

Who? unfoo

That's right, another domain for another project: unfoo.com. It probably will not work for a few days. This is where I'm hosting that generalized version of fanconcert I mentioned above. Why the name unfoo? Why not? It was available, it's short and makes about as much sense as calling something a wiki.

But ahhh, it actually is like a wiki in a way. unfoo will be a general framework to hold objects that can be edited collaboratively by a whole lot of people. Wiki's work this way, except the data on a wiki is prose.

Prose is not very useful to a computer because it's hard for computers to understand a piece of prose's semantics. Computers can index prose by keeping track of where certain keywords are and that's what search engines do. If the information were more organized, computers would be able to do a lot more with it.

That's what I want to do with unfoo (and eventually fanconcert): let users (and bots) enter information in a structured manner. The structure will be familiar to programmers: objects with types (classes), attributes and relationships with other objects. unfoo will be a human usable website and an API to add and modify these objects and their attributes and store them so the objects can be used by other websites (like fanconcert).

It would also be interesting to release unfoo into the wild like wikis, get other people to set up other unfoos for specific types of information (domains) and then link several unfoos together over the Internet to form a sort of collective intelligence. When I hear the phrase "Semantic Web" this is what comes to mind ... why hasn't someone built it?